Información del artículo
Fundamentos: En el campo de las enfermedades infecciosas y la microbiología clínica la evaluación de pruebas diagnósticas (PD) es un área de investigación fundamental. Las dificultades específicas de este tipo de estudios han motivado que no haya alcanzado el rigor metodológico de otras áreas de investigación clínica. Este artículo pretende valorar y caracterizar la metodología de los estudios sobre PD publicados en la revista Enfermedades Infecciosas y Microbiología Clínica (EIMC).
Material y métodos: Se seleccionaron 45 artículos publicados en el período 1990-1996 en los que se determinaba la sensibilidad y especificidad de diversas PD. Se aplicó un conjunto de criterios de calidad metodológica ampliamente aceptados.
Resultados: Todos los trabajos salvo uno (98%) especificaban el patrón de referencia empleado, si bien en cuatro (9%) éste incluía la PD (sesgo de incorporación). La descripción adecuada de la PD se recoge en el 75% de los casos, pero sólo en el 11% se evaluaba la reproducibilidad. La especificación de la fuente de la población de referencia, criterios de inclusión y espectro de sujetos/especímenes se presentaron en el 58, 33 y 40% de los artículos, respectivamente. Un 33% de los trabajos presentaban sesgo de secuencia diagnóstica, sólo un 6% especificaban el análisis ciego de los resultados y un 11% tuvieron en cuenta los posibles resultados indeterminados de la PD. La mitad de los trabajos estimaban la sensibilidad y especificidad por estratos clínicos, únicamente un trabajo (2%) calculaba la precisión de los índices (intervalos de confianza) y un 7% presentaban curvas de rendimiento diagnóstico.
Conclusión: La calidad metodológica de la investigación sobre PD recogidas en la revista EIMC es mejorable en el apartado de diseño y presentación de resultados.
Palabras clave: Pruebas diagnósticas. Microbiología.
Methods of evaluation of diagnostic tests in Enfermedades Infecciosas y Microbiología Clínica
Background: In the field of infectious diseases and clinical microbiology, the evaluation of diagnostic tests (DT) is an important research area. The specific difficulties of this type of research has motivated that have not caught the severity methodological of others areas of clinical research. This article try to asses and characterize the methodology of articles about DT published in Enfermedades Infecciosas y Microbiología Clínica (EIMC) journal.
Methods: Forty-five articles was selected in the EIMC journal during the 1990-1996 period, because of determinate the sensitivity and specificity of different DT. Methodological standards, extensively accepted was used.
Results: In all of articles, except one (98%) the gold standard was specified yours use, however in 4 studies (9%) include the DT in the gold standard (incorporation bias). The correct description of DT was reported in 75% of cases, but only in 11% cases the reproducibility of test was evaluated. The description of source of reference population, standard of inclusion and spectrum of composition was described in 58, 33 and 40% of articles, respectively. In 33% of studies presented workup bias, only 6% commented blind-analysis of results, and 11% presented indeterminate test results. Half of the studies reported test indexes for clinical subgroups, only one article (2%) provided numerical precision for test indexes, and only 7% reported receiver operating characteristics curves.
Conclusions: The methodological quality of DT research in the EIMC journal may improve in different aspects of design and presentation of results
Material y métodos: Se seleccionaron 45 artículos publicados en el período 1990-1996 en los que se determinaba la sensibilidad y especificidad de diversas PD. Se aplicó un conjunto de criterios de calidad metodológica ampliamente aceptados.
Resultados: Todos los trabajos salvo uno (98%) especificaban el patrón de referencia empleado, si bien en cuatro (9%) éste incluía la PD (sesgo de incorporación). La descripción adecuada de la PD se recoge en el 75% de los casos, pero sólo en el 11% se evaluaba la reproducibilidad. La especificación de la fuente de la población de referencia, criterios de inclusión y espectro de sujetos/especímenes se presentaron en el 58, 33 y 40% de los artículos, respectivamente. Un 33% de los trabajos presentaban sesgo de secuencia diagnóstica, sólo un 6% especificaban el análisis ciego de los resultados y un 11% tuvieron en cuenta los posibles resultados indeterminados de la PD. La mitad de los trabajos estimaban la sensibilidad y especificidad por estratos clínicos, únicamente un trabajo (2%) calculaba la precisión de los índices (intervalos de confianza) y un 7% presentaban curvas de rendimiento diagnóstico.
Conclusión: La calidad metodológica de la investigación sobre PD recogidas en la revista EIMC es mejorable en el apartado de diseño y presentación de resultados.
Palabras clave: Pruebas diagnósticas. Microbiología.
Methods of evaluation of diagnostic tests in Enfermedades Infecciosas y Microbiología Clínica
Background: In the field of infectious diseases and clinical microbiology, the evaluation of diagnostic tests (DT) is an important research area. The specific difficulties of this type of research has motivated that have not caught the severity methodological of others areas of clinical research. This article try to asses and characterize the methodology of articles about DT published in Enfermedades Infecciosas y Microbiología Clínica (EIMC) journal.
Methods: Forty-five articles was selected in the EIMC journal during the 1990-1996 period, because of determinate the sensitivity and specificity of different DT. Methodological standards, extensively accepted was used.
Results: In all of articles, except one (98%) the gold standard was specified yours use, however in 4 studies (9%) include the DT in the gold standard (incorporation bias). The correct description of DT was reported in 75% of cases, but only in 11% cases the reproducibility of test was evaluated. The description of source of reference population, standard of inclusion and spectrum of composition was described in 58, 33 and 40% of articles, respectively. In 33% of studies presented workup bias, only 6% commented blind-analysis of results, and 11% presented indeterminate test results. Half of the studies reported test indexes for clinical subgroups, only one article (2%) provided numerical precision for test indexes, and only 7% reported receiver operating characteristics curves.
Conclusions: The methodological quality of DT research in the EIMC journal may improve in different aspects of design and presentation of results
Palabras clave:
Diagnostic test
Microbiology
Background: In the field of infectious diseases and clinical microbiology, the evaluation of diagnostic tests (DT) is an important research area. The specific difficulties of this type of research has motivated that have not caught the severity methodological of others areas of clinical research. This article try to asses and characterize the methodology of articles about DT published in Enfermedades Infecciosas y Microbiología Clínica (EIMC) journal.
Methods: Forty-five articles was selected in the EIMC journal during the 1990-1996 period, because of determinate the sensitivity and specificity of different DT. Methodological standards, extensively accepted was used.
Results: In all of articles, except one (98%) the gold standard was specified yours use, however in 4 studies (9%) include the DT in the gold standard (incorporation bias). The correct description of DT was reported in 75% of cases, but only in 11% cases the reproducibility of test was evaluated. The description of source of reference population, standard of inclusion and spectrum of composition was described in 58, 33 and 40% of articles, respectively. In 33% of studies presented workup bias, only 6% commented blind-analysis of results, and 11% presented indeterminate test results. Half of the studies reported test indexes for clinical subgroups, only one article (2%) provided numerical precision for test indexes, and only 7% reported receiver operating characteristics curves.
Conclusions.: The methodological quality of DT research in the EIMC journal may improve in different aspects of design and presentation of results
Methods: Forty-five articles was selected in the EIMC journal during the 1990-1996 period, because of determinate the sensitivity and specificity of different DT. Methodological standards, extensively accepted was used.
Results: In all of articles, except one (98%) the gold standard was specified yours use, however in 4 studies (9%) include the DT in the gold standard (incorporation bias). The correct description of DT was reported in 75% of cases, but only in 11% cases the reproducibility of test was evaluated. The description of source of reference population, standard of inclusion and spectrum of composition was described in 58, 33 and 40% of articles, respectively. In 33% of studies presented workup bias, only 6% commented blind-analysis of results, and 11% presented indeterminate test results. Half of the studies reported test indexes for clinical subgroups, only one article (2%) provided numerical precision for test indexes, and only 7% reported receiver operating characteristics curves.
Conclusions.: The methodological quality of DT research in the EIMC journal may improve in different aspects of design and presentation of results
Keywords:
Diagnostic test
Microbiology
